Skip to main content

Semisupervised learning of hidden Markov models via a homotopy method.

Publication ,  Journal Article
Ji, S; Watson, LT; Carin, L
Published in: IEEE transactions on pattern analysis and machine intelligence
February 2009

Hidden Markov model (HMM) classifier design is considered for the analysis of sequential data, incorporating both labeled and unlabeled data for training; the balance between the use of labeled and unlabeled data is controlled by an allocation parameter \lambda \in [0, 1), where \lambda = 0 corresponds to purely supervised HMM learning (based only on the labeled data) and \lambda = 1 corresponds to unsupervised HMM-based clustering (based only on the unlabeled data). The associated estimation problem can typically be reduced to solving a set of fixed-point equations in the form of a "natural-parameter homotopy." This paper applies a homotopy method to track a continuous path of solutions, starting from a local supervised solution (\lambda = 0) to a local unsupervised solution (\lambda = 1). The homotopy method is guaranteed to track with probability one from \lambda = 0 to \lambda = 1 if the \lambda = 0 solution is unique; this condition is not satisfied for the HMM since the maximum likelihood supervised solution (\lambda = 0) is characterized by many local optima. A modified form of the homotopy map for HMMs assures a track from \lambda = 0 to \lambda = 1. Following this track leads to a formulation for selecting \lambda \in [0, 1) for a semisupervised solution and it also provides a tool for selection from among multiple local-optimal supervised solutions. The results of applying the proposed method to measured and synthetic sequential data verify its robustness and feasibility compared to the conventional EM approach for semisupervised HMM training.

Duke Scholars

Published In

IEEE transactions on pattern analysis and machine intelligence

DOI

EISSN

1939-3539

ISSN

0162-8828

Publication Date

February 2009

Volume

31

Issue

2

Start / End Page

275 / 287

Related Subject Headings

  • Pattern Recognition, Automated
  • Models, Theoretical
  • Models, Statistical
  • Markov Chains
  • Data Interpretation, Statistical
  • Computer Simulation
  • Artificial Intelligence & Image Processing
  • Artificial Intelligence
  • Algorithms
  • 4611 Machine learning
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Ji, S., Watson, L. T., & Carin, L. (2009). Semisupervised learning of hidden Markov models via a homotopy method. IEEE Transactions on Pattern Analysis and Machine Intelligence, 31(2), 275–287. https://doi.org/10.1109/tpami.2008.71
Ji, Shihao, Layne T. Watson, and Lawrence Carin. “Semisupervised learning of hidden Markov models via a homotopy method.IEEE Transactions on Pattern Analysis and Machine Intelligence 31, no. 2 (February 2009): 275–87. https://doi.org/10.1109/tpami.2008.71.
Ji S, Watson LT, Carin L. Semisupervised learning of hidden Markov models via a homotopy method. IEEE transactions on pattern analysis and machine intelligence. 2009 Feb;31(2):275–87.
Ji, Shihao, et al. “Semisupervised learning of hidden Markov models via a homotopy method.IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 31, no. 2, Feb. 2009, pp. 275–87. Epmc, doi:10.1109/tpami.2008.71.
Ji S, Watson LT, Carin L. Semisupervised learning of hidden Markov models via a homotopy method. IEEE transactions on pattern analysis and machine intelligence. 2009 Feb;31(2):275–287.

Published In

IEEE transactions on pattern analysis and machine intelligence

DOI

EISSN

1939-3539

ISSN

0162-8828

Publication Date

February 2009

Volume

31

Issue

2

Start / End Page

275 / 287

Related Subject Headings

  • Pattern Recognition, Automated
  • Models, Theoretical
  • Models, Statistical
  • Markov Chains
  • Data Interpretation, Statistical
  • Computer Simulation
  • Artificial Intelligence & Image Processing
  • Artificial Intelligence
  • Algorithms
  • 4611 Machine learning